How about Dynamic range?

[Canochrome]

I had a 40D for about a year, and used it alongside my 5D. I didn't see it as being all that close to the 5D. If you hit exposure, it was fine. But when I was shooting ISO 400 or higher and underexposed by even 1/3 stop, I saw a significant increase in noise as I pulled up the exposure to make it correct. I didn't see that with my 30D. Although it wasn't quite up to the quality of the 5D, it did have a little more laittude than the 40D did.

The 7D seems to be better than the 40D in that respect, though I've never shot the two side-by-side. It is the closest to the 5D that I've seen from a crop camera.

Now, the 5D2 is even better especially at higher ISO settings.

Look I did a lot of shooting w different cameras in Raw only and processed in DPP and CS3 and AS SOON as you open the 40d file there's no need for peeping and tweaking..you can see it right away! my findings were that the 40d is better than the 50d (specially the shadows) WAY BETTER than the 30d,20d,XT (in highlights recovery also...40d can recover more highlight detail).

That said, 5d files were better than the 40d in DR (slightly)...and so were MkIII files but that is to be expected really given the sensors

Note: Before you jump on me I want to add that i'm a semi-professional photographer...and yes, I sell quite a lot of pics and I have a free-lance contract w a luxury magazine. These were MY FINDINGS shooting anything from events,jewellery and mostly wildlife.

--
Tom

 

[Canochrome]

That's probably a better way of wording it. I almost routinely find at least 1 stop worth of recoverable highlights in a RAW image that I can capture by adjusting the exposure. The camera's captured information that will appear blown in the JPG but can be found in the original RAW image. Recurving the exposure can give you this detail without hiding shadow detail. This is why people say that RAW offers greater dynamic range than JPG images.

I think Canon offers HTP for people that, for various reasons, don't shoot RAW or have to use JPG 'out-of-the-camera' to meet deadlines. The camera basically does the work of recovering that stop or so of highlight detail that would normally be lost in a 'straight-from-the-camera' JPG image. I can see the advantage if you don't want to process RAW images for some reason or another.

Sports shooters often use JPG because of time constraints. In fact, many use 1-series cameras and record the JPG on one card and the RAW on the other so that they have the higher-quality RAW image to use if they later need to work with it a bit.

OK, then let's say that RAW has more 'headroom' than JPEG (is that OK?), but why is that? Why don't the in-camera JPEGs take full advantage of all the information and DR that's available in the original RAW capture? Why does Canon for example use a mode like HTP (which just is a 1 stop under-exposure, resulting in more noise) to prevent blown highlights, if there's already 1 stop highlight headroom available in the RAW file? Isn't that a bit silly!?

--
Tom

 

[rhlpetrus]

No, RAW has no "headroom". RAW file is what it is, there's clipping there or not. What may extract more or less of that is the conversion from RAW to other formats.

Supposedly DXOMArk measures RAW DR, but there's some controversy about if their testing is well-designed and done correctly. Anyway, it's clear it's close to what one expect, MF cameras do better, then FF, then APS-C, them 4:3, then compacts.

Some softwares have better performance extracting DR than others, and some have recovery tools, meaning essentially that they change the top and bottom curve tones to do that (typical jpeg conversion tone curve gives this 8-9stop DR values DPR tests.

Best APS-C sensor measured there is D90/D5000 but Gabor has shown some trick done at RAW level may be causing that, not sure.

Anyway, 12 stops is already pretty good if one can really get all that from sensor, better already than positive film and similar to negative film.
--
Renato.
http://www.flickr.com/photos/rhlpedrosa/
OnExposure member
http://www.onexposure.net/

Good shooting and good luck
(after Ed Murrow)

 

[GaborSch]

OK, then let's say that RAW has more 'headroom' than JPEG (is that OK?), but why is that?

The metering is the same, no matter of raw and/or JPEG output is selected. Accordingly, the exposure, clipping is the same. Now, what is the difference, what can be "gained" by raw, what is that "headroom"?

The standard picture styles include an "S" curve", which makes the very highlights uncontrasty; this may be interpreted as losing highlights. One can create a "linear" picture style (I am using one for better ETTR judgement); but that would be detrimental to the overal appearance of an in-camera JPEG, for it reduces the contrast in the mid-range.

--
Gabor

http://www.panopeeper.com/panorama/pano.htm

 

[Adam2]

I took my 7d and 1dmkII out in the woods today around sunset. While the 7d images looked fairly good on the camera display I was quite disappointed once they were viewed on the monitor. Although I shot in RAW and processed in DPP, there was something amiss in the highlights of the 7d images. They appeared a bit clipped though the rest of the image was properly exposed. Compared to the venerable 1dmkII, the differences were striking.

For comparison, the 1dmkII has around 9.2 stops of DR to the 7d's 8.3. Also, it appears from the review that the 7d exhibits less DR headroom in the highlights. No amount of PP could rectify the rather glaring distinctions. Personally, for the $ I would pick up a 5d or 1dmkII (if better AF is needed) rather than a 7d. Alternatively, wait a couple of months and the price of a good used 1dmkIII should be around the same $. While I loved the high ISO performance and improved resolution, and appreciated the improved AF compared to other xxd bodies, the DR issue is a disappointment. The 7d is going back.

 

[Steen Bay]

OK, then let's say that RAW has more 'headroom' than JPEG (is that OK?), but why is that?

The metering is the same, no matter of raw and/or JPEG output is selected. Accordingly, the exposure, clipping is the same. Now, what is the difference, what can be "gained" by raw, what is that "headroom"?

The standard picture styles include an "S" curve", which makes the very highlights uncontrasty; this may be interpreted as losing highlights. One can create a "linear" picture style (I am using one for better ETTR judgement); but that would be detrimental to the overal appearance of an in-camera JPEG, for it reduces the contrast in the mid-range.

Are you saying that RAW and in-camera JPEG actually are 'clipping' at the same point? That there isn't any additional highlight information (or 'headroom') in RAW? That it's just the tone curve used that's compressing the highlights in the JPEG? If so, is it then also possible to 'recover' (sorry ;-)) the highlights from a in-camera JPEG? (Or maybe only if a custom 'linear' picture style is used?)

 

[Peter 13]

...

Higher DR in JPEG is not necessarily superiority. Your medium (monitor/paper) has a fixed DR, regardless of the camera. To fit higher DR of a scene on it, you have to compress it more than usual. That can make the image look flat.
...

Look flat, yes as sometimes happens with HDR images. But one always has the option of discarding DR in processing to restore the "photographic" look (low dr) to a print. Unfortunately, there is no way to create DR that was not captured in the sensor. For my money, higher DR at capture is always a plus.

Higher DR of the JPEG does not mean higher DR sensor. What I am trying to say is that DPR measures the DR of processed (in some particular way) images.

 

[GaborSch]

Are you saying that RAW and in-camera JPEG actually are 'clipping' at the same point?

Exactly. I tested this: I made several shots, with increasing exposure, RAW+JPEG. I saw the clipped spots in Rawnalyze - and the JPEG containes 255 on the same locations, but not elsewhere.

That there isn't any additional highlight information (or 'headroom') in RAW?

That's not the same claim. The JPEG data is mapped in the respective color space; that implies the nonlinear transformation as well ("gamma curve"). That includes the loss of lots of original pixel values. Even more: the applying the S-curve eliminates most of the original pixel values from the very highlights and shadows. The RGB values 253, 254, 255 (only three different levels) may replace dozens of the mapped values, which os hundreds of raw pixel values.

You can reduce the intensity ("exposure" in ACR) in the raw conversion, or use "Recovery" (in ACR) or some custom curve, in order to assign more output levels to the highlights. Suddenly details of the clouds appear in the formerly plain white sky or on bright cloths, etc.

That it's just the tone curve used that's compressing the highlights in the JPEG? If so, is it then also possible to 'recover' (sorry ;-)) the highlights from a in-camera JPEG?

That can not be recovered from JPEG any more.

Or maybe only if a custom 'linear' picture style is used?

That is possible, but impractical: you would have to know before making the shot , if to use this. Using such a curve is double edged: if you "spend" more levels on the highlights, you have less levels for the rest. Do you know this before having analyzed the shot?

On the other hand, one can not only analyze and decide this during conversion of the raw data, but one has more options: a special curve for that shot to distribute the levels differently, or even developing two (or more) results from the same raw and blanding them: a fake HDR.

--
Gabor

http://www.panopeeper.com/panorama/pano.htm

 

[GaborSch]

How many cameras have you used IN THE FIELD. How many shoots you covered w these cameras?

I still have a 20D and a 40D, a few ten thousand shots. However, this is irrelevant. I don't have a 7D, 5D2, Sony A900, Nikon this and that, but I know more about them in certain aspects than the vast majority of the owners.

So enlighten us w all your work published in actual magazines (given ur such a smart-a$$ / know it all guy I assume magazines are dyin' to pay for ur work) or share your great work as a photographer and maybe throw in some pics you got paid for. If not, shut up and crowl back into ur "behind the keyboard " hole dude

What you don't understand is, that the current topic is not your general ability as photographer. You may be a good photographer, I did not say anything regarding that. However, your understanding of certain important aspects of digital photography are on pretty low level for working with raw files .

--
Gabor

http://www.panopeeper.com/panorama/pano.htm

 

[JerryG1]

...

Higher DR of the JPEG does not mean higher DR sensor. What I ma trying to say is that DPR measures the DR of processed (in some particular way) images.

I agree with that.

Still, if DR difference is due entirely to the raw to jpeg conversion, I wonder why Canon decided to not continue the method of the 40D which led to higher DR.
--
JerryG

My galleries at:
http://www.pbase.com/jerryg1

 

[rondhamalam]

Agree

40D has a lot of room to play the RAW with.

Even on Jpeg test, DPR stated that 40D has more dynamic range than D300 and 50D.

And I have the same experience on 40D, compare to my 5D2, the 40D has wider dynamic range.
(Especially when 5D2 has vertical banding problem at low iso on shadows area)

The value of a camera is when photographer is starting to post-process the RAW file.

Look I did a lot of shooting w different cameras in Raw only and processed in DPP and CS3 and AS SOON as you open the 40d file there's no need for peeping and tweaking..you can see it right away! my findings were that the 40d is better than the 50d (specially the shadows) WAY BETTER than the 30d,20d,XT (in highlights recovery also...40d can recover more highlight detail).

That said, 5d files were better than the 40d in DR (slightly)...and so were MkIII files but that is to be expected really given the sensors

Note: Before you jump on me I want to add that i'm a semi-professional photographer...and yes, I sell quite a lot of pics and I have a free-lance contract w a luxury magazine. These were MY FINDINGS shooting anything from events,jewellery and mostly wildlife.

 

[Steen Bay]

Are you saying that RAW and in-camera JPEG actually are 'clipping' at the same point?

Exactly. I tested this: I made several shots, with increasing exposure, RAW+JPEG. I saw the clipped spots in Rawnalyze - and the JPEG containes 255 on the same locations, but not elsewhere.

That there isn't any additional highlight information (or 'headroom') in RAW?

That's not the same claim. The JPEG data is mapped in the respective color space; that implies the nonlinear transformation as well ("gamma curve"). That includes the loss of lots of original pixel values. Even more: the applying the S-curve eliminates most of the original pixel values from the very highlights and shadows. The RGB values 253, 254, 255 (only three different levels) may replace dozens of the mapped values, which os hundreds of raw pixel values.

You can reduce the intensity ("exposure" in ACR) in the raw conversion, or use "Recovery" (in ACR) or some custom curve, in order to assign more output levels to the highlights. Suddenly details of the clouds appear in the formerly plain white sky or on bright cloths, etc.

That it's just the tone curve used that's compressing the highlights in the JPEG? If so, is it then also possible to 'recover' (sorry ;-)) the highlights from a in-camera JPEG?

That can not be recovered from JPEG any more.

Or maybe only if a custom 'linear' picture style is used?

That is possible, but impractical: you would have to know before making the shot , if to use this. Using such a curve is double edged: if you "spend" more levels on the highlights, you have less levels for the rest. Do you know this before having analyzed the shot?

On the other hand, one can not only analyze and decide this during conversion of the raw data, but one has more options: a special curve for that shot to distribute the levels differently, or even developing two (or more) results from the same raw and blanding them: a fake HDR.

OK, I can see your point. The problem is that there's only 256 tone values available in the 8 bit JPEG, and a (more) linear tone curve would give us a very 'flat' image. But let's go back to HTP: A ISO 200 HTP image is actually an underexposed ISO 100 image, in which the shadows/midtones are 'lifted' 1 stop, resulting in more noise than a correctly exposed ISO 100 image would have had.

Wouldn't it be possible (and much better) to simply use another (in-camera) highlight protecting tone curve on the correctly exposed ISO 100 image instead, and thus take advantage of the 'headroom' (or whatever..) that's already (in practice) available in the original 12/14 bit capture? Or in other words: isn't it possible to 'recover' the same amount of highlight detail/information from a correctly exposed ISO 100 RAW image as the more noisy ISO 200 HTP JPEG image (which actually is a 1 stop underexposed ISO 100 image) gives us?

 

[RedFox88]

DR per image size is not the same as DR per pixel. According to DXOmark, it is 11+ stops for both the 50D and the 40D, with some advantage for the 50D. I do not see anything exceptional about the 40D in that regard.

That is one website. Yet DPR gives the 40D 1.2 stops more DR than the 7D.

Again, this is DR of the converted JPEG. Give me a RAW file from any recent dSLR and I can produce a JPEG with that DR in a few seconds.

Can you read? It's called RAW headroom, not JPG headroom. :O

 

[GaborSch]

Wouldn't it be possible (and much better) to simply use another (in-camera) highlight protecting tone curve on the correctly exposed ISO 100 image instead, and thus take advantage of the 'headroom' (or whatever..) that's already (in practice) available in the original 12/14 bit capture? Or in other words: isn't it possible to 'recover' the same amount of highlight detail/information from a correctly exposed ISO 100 RAW image as the more noisy ISO 200 HTP JPEG image (which actually is a 1 stop underexposed ISO 100 image) gives us?

This would be possible, but it could save only that, what was not clipped in the shot but rendered "flat" by the in-camera raw conversion. The underexposure caused by HTP is supposed to save the shot from being clipped .

HTP should never be used with raw. One full stop lower exposure is simply brutal; we can use -1/3 EV or -2/3 EV exposure bias, when it seems necessary. However, I accept, that it is reasonable with JPEG, for

1. clandestine changing the exposure , i.e. shutter or aperture by the camera is IMO unacceptable, for those settings are part of the "shot design" (except in P mode), while changing the ISO is a pure technical issue,

2. changing the ISO by 1/3 stop is useless, except with the 1Dxxx cameras, because there is no 1/3 stop ISO, it's an eyewash.

For JPEG shooters a new option would be useful: if there is an exposure bias, not more than let's say 1 EV, then lift the shadows by so much if the bias is positive or reduce the highlights gradually if the bias is negative. However, I realize that the mass of options is already a pain for many users. I regularly see on forums, that most users don't understand the effect of HTP, nor of fake ISOs, etc.

--
Gabor

http://www.panopeeper.com/panorama/pano.htm

 

[Peter 13]

DR per image size is not the same as DR per pixel. According to DXOmark, it is 11+ stops for both the 50D and the 40D, with some advantage for the 50D. I do not see anything exceptional about the 40D in that regard.

That is one website. Yet DPR gives the 40D 1.2 stops more DR than the 7D.

Again, this is DR of the converted JPEG. Give me a RAW file from any recent dSLR and I can produce a JPEG with that DR in a few seconds.

Can you read? It's called RAW headroom, not JPG headroom. :O

WTF are you talking about?

 

[Mark H]

DR per image size is not the same as DR per pixel. According to DXOmark, it is 11+ stops for both the 50D and the 40D, with some advantage for the 50D. I do not see anything exceptional about the 40D in that regard.

That is one website. Yet DPR gives the 40D 1.2 stops more DR than the 7D.

Again, this is DR of the converted JPEG. Give me a RAW file from any recent dSLR and I can produce a JPEG with that DR in a few seconds.

Can you read? It's called RAW headroom, not JPG headroom. :O

WTF are you talking about?

Less of the "WTF" - and maybe try paying more attention.

In case you hadn't noticed - the title of all the messages you have replied to refers to RAW - not JPEG.

The information 'RedFoxx88' referred to was DPR's own conversion of RAW format test images using ACR - not JPEG.

 

[Peter 13]

DR per image size is not the same as DR per pixel. According to DXOmark, it is 11+ stops for both the 50D and the 40D, with some advantage for the 50D. I do not see anything exceptional about the 40D in that regard.

That is one website. Yet DPR gives the 40D 1.2 stops more DR than the 7D.

Again, this is DR of the converted JPEG. Give me a RAW file from any recent dSLR and I can produce a JPEG with that DR in a few seconds.

Can you read? It's called RAW headroom, not JPG headroom. :O

WTF are you talking about?

Less of the "WTF" - and maybe try paying more attention.

In case you hadn't noticed - the title of all the messages you have replied to refers to RAW - not JPEG.

The information 'RedFoxx88' referred to was DPR's own conversion of RAW format test images using ACR - not JPEG.

I agree. The title often stays unchanged while the discussion takes a different turn but going a few messages up makes it clear.

As far as the "WTF" is converned - it was not me who changed the tone first.

 

[Mark H]

No, RAW has no "headroom". RAW file is what it is, there's clipping there or not. What may extract more or less of that is the conversion from RAW to other formats.

The term 'headroom' , in this context, refers to the information in the RAW data for a given exposure, compared to that of a JPEG (or similar processed format) of the same given exposure.

'Simples'.

 

[RedFox88]

The information 'RedFoxx88' referred to was DPR's own conversion of RAW format test images using ACR - not JPEG.

I agree. The title often stays unchanged while the discussion takes a different turn but going a few messages up makes it clear.

As far as the "WTF" is converned - it was not me who changed the tone first.

I change the subject on most of my replies, because I used threaded view and absolutely hate looking at a post with 30 replies and they all are unchanged subject replies.

DPR "gave up" on encouraging repliers to change the subject line when they switched the starting cursor on replies to the body area instead of the subject line. They encourage repliers to change the subject line to describe your response or to start your response.

 

[Peter 13]

Sorry for the overreaction, I did not read the previous posts carefully.